A Buffering Approach to Manage I/O in a Normalized Cross-Correlation Earthquake Detection Code for Large Seismic Datasets

D
A Buffering Approach to Manage I/O in a
Normalized Cross-Correlation Earthquake
Detection Code for Large Seismic Datasets
Dawei Mu, Pietro Cicotti, Yifeng Cui, Enjui Lee, Po Chen
Outlines
1. Introduction of cuNCC code
2. Realistic Application
3. Performance Analysis
4. Memory Buffer Approach and I/O Analysis
5. Future Work
1. Introduction of cuNCC
what is cuNCC ?
CUDA based software designed to calculate the normalized cross-
correlation coefficient (NCC) between a collection of selected
template waveforms and the continuous waveform recordings of
seismic instruments to evaluate the waveform similarity among the
waveforms and/or the relative travel-time differences.
Feb 05, 2016 M6.6

Meinong aftershocks detection
• more uncatalogued aftershocks

were detected
• contributed to earthquake 

location detection and earthquake 

source parameters estimation
2. Realistic Application
M6.6 Meinong aftershocks 

hypocenter re-location.
•traditional method using short term 

long term to detect events and using 

1-D model to locate hypocenter. fewer

aftershock detected, the result contains

less information due to inaccuracy.
•3-D waveform template matching detect

events, using 3D model and waveform 

travel-time differences to re-locate the 

hypocenter, the result shows more 

events and more clustered hypocenters, 

which give us detailed fault geometry
•over 4 trillion NCC calculations involved
3. Performance Analysis
optimization scheme
• The cuNCC is bounded by 

memory bandwidth
• The constant memory is used to 

stack multiple events into 

single computational kernel
• The shared memory is used to improve the memory bandwidth 

utilization
1. Compute, Bandwidth, or Latency Bound
The first step in analyzing an individual kernel is to determine if the performance of the kernel is bounded by com
bandwidth, or instruction/memory latency. The results below indicate that the performance of kernel "cuNCC_04
limited by memory bandwidth. You should first examine the information in the "Memory Bandwidth" section to d
limiting performance.
1.1. Kernel Performance Is Bound By Memory Bandwidth
For device "GeForce GTX 980" the kernel's compute utilization is significantly lower than its memory utilization
levels indicate that the performance of the kernel is most likely being limited by the memory system. For this kern
factor in the memory system is the bandwidth of the Shared memory.
3. Performance Analysis
performance benchmark
This cuNCC code can achieve high performance without high-end
hardware or expensive clusters, optimized CPU based NCC code
needs 21 hours for one E7-8867 CPU (all 18 cores) to finish
mentioned example while a NVIDIA GTX980 only costs 53 minutes.
Hardware Runtime (ms)
SP FLOP
(×1e11)
Achieved
GFLOPS
Max GFLOPS
Achieved

Percentage
Speedup
E7-8867

(18 cores)
2968 1.23 41.36 237.6 17.4% 1.0x
C2075

(Fermi)
495 1.8 363.83 1030 35.3% 6.0x
GTX980

(Maxwell)
116 1.8 1552.80 5000 31.0% 25.6x
M40

(Maxwell)
115 1.8 1569.86 7000 22.4% 25.8x
P100

(Pascal)
62 1.8 2911.84 10600 27.5% 47.9x
4. Memory Buffer Approach and I/O Analysis
the I/O bottleneck
After improving the computational 

performance with GPU acceleration, 

I/O efficiency became the new 

bottleneck of the cuNCC’s overall 

performance.
The output file of cuNCC is an 1-D vector of similarity coefficients
saved in binary format, which size is equal to seismic data file.
CPU NCC I/O operations cost roughly 10% of the total runtime, while
the GPU code I/O cost more than 75% of total runtime.
0
125
250
375
500
NCC(CPU) cuNCC
I/O Compute
4. Memory Buffer Approach and I/O Analysis
test environment
The SGI UV300 system has 8 sockets 18 core Intel Xeon E7-8867 V4
processors 16 DDR4 32GB DRAM run at 1600 MHz.
4 TB of DRAM in Non-Uniform Memory Access (NUMA) via SGI’s
NUMALink technology.
4x PCIe Intel flash cards for a total of 8TB configured as a RAID 0
device and mounted as “/scratch” with a 975.22 MB/s achieved I/O
bandwidth (with IOR).
2x 400GB Intel SSDs configured as a RAID 1 device and mounted as
“/home” with a 370.07 MB/s achieved I/O bandwidth.
The software we used were GCC-4.4.7 and CUDA-7.5 along with MPI
package MPICH-3.2.
4. Memory Buffer Approach and I/O Analysis
use CPU memory as a buffer
•Most GPU-enabled computers have more CPU

memory than GPU memory. 

(in our case 48GB << 4TB)
•Fixed data chuck size (120 days’) 

with different total workloads
•on ”/scratch” partition, for every data size, the 

buffering technique costs more overall runtime 

than the no-buffering
•on ”/home” partition, buffering starts to help 

after reaching the 2400-day’s total workload
•the high I/O bandwidth filesystem, the 

improvement brought by the buffering cannot 

cover up the overhead of the memory transfer.
4. Memory Buffer Approach and I/O Analysis
use shared memory virtual 

filesystem as a buffer
•we set 2 TB of DRAM as a shared memory 

virtual filesystem, and measured the I/O 

bandwidth achieved 2228.05 MB/s.
•on the ”/dev/shm” partition, the high 

bandwidth of shared memory improves 

performance greatly by reducing the 

time used for output.
•we gathered the runtime result without 

buffering scheme from all three storage 

partitions, and the shared memory 

partition obtains the best performance.
4. Memory Buffer Approach and I/O Analysis
I/O test conclusion
•for machines support shared memory virtual filesystem, we
recommend to use the shared memory as buffer to output cuNCC
result, especially when the similarity coefficients are the median
result for the following computation.
•for those machines do not have shared memory with high bandwidth
I/O device, we recommend to directly output the result to storage
without the buffering scheme.
•for those machines do not support shared memory with low
bandwidth I/O device, we should consider to use CPU memory as a
buffer to reduce disk access frequency.
5. Future Work
• further optimize the cuNCC code on the Pascal GPU platform.
• implement our cuNCC code with “SEISM-IO” library, which interface
allows user to switch among “MPI-IO”, “PHDF5”, “NETCDF4”, and
“ADIOS” as low level I/O libraries.
Thank you for your time !
1 of 13

Recommended

USENIX NSDI 2016 (Session: Resource Sharing) by
USENIX NSDI 2016 (Session: Resource Sharing)USENIX NSDI 2016 (Session: Resource Sharing)
USENIX NSDI 2016 (Session: Resource Sharing)Ryousei Takano
1.1K views16 slides
クラウド時代の半導体メモリー技術 by
クラウド時代の半導体メモリー技術クラウド時代の半導体メモリー技術
クラウド時代の半導体メモリー技術Ryousei Takano
2.5K views24 slides
Applying Recursive Temporal Blocking for Stencil Computations to Deeper Memor... by
Applying Recursive Temporal Blocking for Stencil Computations to Deeper Memor...Applying Recursive Temporal Blocking for Stencil Computations to Deeper Memor...
Applying Recursive Temporal Blocking for Stencil Computations to Deeper Memor...Tokyo Institute of Technology
345 views19 slides
Programming Trends in High Performance Computing by
Programming Trends in High Performance ComputingProgramming Trends in High Performance Computing
Programming Trends in High Performance ComputingJuris Vencels
437 views16 slides
Exploring the Performance Impact of Virtualization on an HPC Cloud by
Exploring the Performance Impact of Virtualization on an HPC CloudExploring the Performance Impact of Virtualization on an HPC Cloud
Exploring the Performance Impact of Virtualization on an HPC CloudRyousei Takano
9.8K views21 slides
Gossip-based resource allocation for green computing in large clouds by
Gossip-based resource allocation for green computing in large cloudsGossip-based resource allocation for green computing in large clouds
Gossip-based resource allocation for green computing in large cloudsRerngvit Yanggratoke
734 views21 slides

More Related Content

What's hot

Final Thesis by
Final ThesisFinal Thesis
Final ThesisDaniel Kosalla
526 views37 slides
User-space Network Processing by
User-space Network ProcessingUser-space Network Processing
User-space Network ProcessingRyousei Takano
5.7K views22 slides
MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola... by
MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola...MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola...
MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola...Heechul Yun
5.9K views29 slides
XNAT Tuning & Monitoring by
XNAT Tuning & MonitoringXNAT Tuning & Monitoring
XNAT Tuning & MonitoringJohn Paulett
1.1K views36 slides
Inp tooptimmempropslinkedinpost 25mar18_004 by
Inp tooptimmempropslinkedinpost 25mar18_004Inp tooptimmempropslinkedinpost 25mar18_004
Inp tooptimmempropslinkedinpost 25mar18_004Michael (Mike) Harwood, PE, PMP
380 views1 slide
Stop-the-world GCs on milticores by
Stop-the-world GCs on milticoresStop-the-world GCs on milticores
Stop-the-world GCs on milticoresAliya Ibragimova
364 views14 slides

What's hot(20)

User-space Network Processing by Ryousei Takano
User-space Network ProcessingUser-space Network Processing
User-space Network Processing
Ryousei Takano5.7K views
MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola... by Heechul Yun
MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola...MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola...
MemGuard: Memory Bandwidth Reservation System for Efficient Performance Isola...
Heechul Yun5.9K views
XNAT Tuning & Monitoring by John Paulett
XNAT Tuning & MonitoringXNAT Tuning & Monitoring
XNAT Tuning & Monitoring
John Paulett1.1K views
AES encryption on modern consumer architectures by Grigore Lupescu
AES encryption on modern consumer architecturesAES encryption on modern consumer architectures
AES encryption on modern consumer architectures
Grigore Lupescu953 views
On heap cache vs off-heap cache by rgrebski
On heap cache vs off-heap cacheOn heap cache vs off-heap cache
On heap cache vs off-heap cache
rgrebski5.7K views
유연하고 확장성 있는 빅데이터 처리 by NAVER D2
유연하고 확장성 있는 빅데이터 처리유연하고 확장성 있는 빅데이터 처리
유연하고 확장성 있는 빅데이터 처리
NAVER D25.2K views
Resource Management with Systemd and cgroups by Tsung-en Hsiao
Resource Management with Systemd and cgroupsResource Management with Systemd and cgroups
Resource Management with Systemd and cgroups
Tsung-en Hsiao362 views
Technology Updates of PG-Strom at Aug-2014 (PGUnconf@Tokyo) by Kohei KaiGai
Technology Updates of PG-Strom at Aug-2014 (PGUnconf@Tokyo)Technology Updates of PG-Strom at Aug-2014 (PGUnconf@Tokyo)
Technology Updates of PG-Strom at Aug-2014 (PGUnconf@Tokyo)
Kohei KaiGai1.5K views
Hioki 8860 memory_hi_corder_datasheet by Angus Sankaran
Hioki 8860 memory_hi_corder_datasheetHioki 8860 memory_hi_corder_datasheet
Hioki 8860 memory_hi_corder_datasheet
Angus Sankaran151 views
Extreme Linux Performance Monitoring and Tuning by Milind Koyande
Extreme Linux Performance Monitoring and TuningExtreme Linux Performance Monitoring and Tuning
Extreme Linux Performance Monitoring and Tuning
Milind Koyande2K views
How Netflix Tunes EC2 Instances for Performance by Brendan Gregg
How Netflix Tunes EC2 Instances for PerformanceHow Netflix Tunes EC2 Instances for Performance
How Netflix Tunes EC2 Instances for Performance
Brendan Gregg524.6K views

Similar to A Buffering Approach to Manage I/O in a Normalized Cross-Correlation Earthquake Detection Code for Large Seismic Datasets

In datacenter performance analysis of a tensor processing unit by
In datacenter performance analysis of a tensor processing unitIn datacenter performance analysis of a tensor processing unit
In datacenter performance analysis of a tensor processing unitJinwon Lee
1.4K views41 slides
Large-Scale Optimization Strategies for Typical HPC Workloads by
Large-Scale Optimization Strategies for Typical HPC WorkloadsLarge-Scale Optimization Strategies for Typical HPC Workloads
Large-Scale Optimization Strategies for Typical HPC Workloadsinside-BigData.com
511 views34 slides
Conference Paper: Universal Node: Towards a high-performance NFV environment by
Conference Paper: Universal Node: Towards a high-performance NFV environmentConference Paper: Universal Node: Towards a high-performance NFV environment
Conference Paper: Universal Node: Towards a high-performance NFV environmentEricsson
791 views7 slides
Kvm performance optimization for ubuntu by
Kvm performance optimization for ubuntuKvm performance optimization for ubuntu
Kvm performance optimization for ubuntuSim Janghoon
40.7K views28 slides
You must create a data file for the project. You can use forma.docx by
You must create a data file for the project. You can use forma.docxYou must create a data file for the project. You can use forma.docx
You must create a data file for the project. You can use forma.docxdavezstarr61655
5 views9 slides
Performance Optimization of CGYRO for Multiscale Turbulence Simulations by
Performance Optimization of CGYRO for Multiscale Turbulence SimulationsPerformance Optimization of CGYRO for Multiscale Turbulence Simulations
Performance Optimization of CGYRO for Multiscale Turbulence SimulationsIgor Sfiligoi
92 views28 slides

Similar to A Buffering Approach to Manage I/O in a Normalized Cross-Correlation Earthquake Detection Code for Large Seismic Datasets(20)

In datacenter performance analysis of a tensor processing unit by Jinwon Lee
In datacenter performance analysis of a tensor processing unitIn datacenter performance analysis of a tensor processing unit
In datacenter performance analysis of a tensor processing unit
Jinwon Lee1.4K views
Large-Scale Optimization Strategies for Typical HPC Workloads by inside-BigData.com
Large-Scale Optimization Strategies for Typical HPC WorkloadsLarge-Scale Optimization Strategies for Typical HPC Workloads
Large-Scale Optimization Strategies for Typical HPC Workloads
inside-BigData.com511 views
Conference Paper: Universal Node: Towards a high-performance NFV environment by Ericsson
Conference Paper: Universal Node: Towards a high-performance NFV environmentConference Paper: Universal Node: Towards a high-performance NFV environment
Conference Paper: Universal Node: Towards a high-performance NFV environment
Ericsson791 views
Kvm performance optimization for ubuntu by Sim Janghoon
Kvm performance optimization for ubuntuKvm performance optimization for ubuntu
Kvm performance optimization for ubuntu
Sim Janghoon40.7K views
You must create a data file for the project. You can use forma.docx by davezstarr61655
You must create a data file for the project. You can use forma.docxYou must create a data file for the project. You can use forma.docx
You must create a data file for the project. You can use forma.docx
davezstarr616555 views
Performance Optimization of CGYRO for Multiscale Turbulence Simulations by Igor Sfiligoi
Performance Optimization of CGYRO for Multiscale Turbulence SimulationsPerformance Optimization of CGYRO for Multiscale Turbulence Simulations
Performance Optimization of CGYRO for Multiscale Turbulence Simulations
Igor Sfiligoi92 views
Morph : a novel accelerator by BaharJV
Morph : a novel acceleratorMorph : a novel accelerator
Morph : a novel accelerator
BaharJV46 views
Approximation techniques used for general purpose algorithms by Sabidur Rahman
Approximation techniques used for general purpose algorithmsApproximation techniques used for general purpose algorithms
Approximation techniques used for general purpose algorithms
Sabidur Rahman476 views
4.1 Introduction 145• In this section, we first take a gander at a.pdf by arpowersarps
4.1 Introduction 145• In this section, we first take a gander at a.pdf4.1 Introduction 145• In this section, we first take a gander at a.pdf
4.1 Introduction 145• In this section, we first take a gander at a.pdf
arpowersarps3 views
Optimizing Performance and Computing Resource Efficiency of In-Memory Big Dat... by Databricks
Optimizing Performance and Computing Resource Efficiency of In-Memory Big Dat...Optimizing Performance and Computing Resource Efficiency of In-Memory Big Dat...
Optimizing Performance and Computing Resource Efficiency of In-Memory Big Dat...
Databricks508 views
Hands-on Lab: How to Unleash Your Storage Performance by Using NVM Express™ B... by Odinot Stanislas
Hands-on Lab: How to Unleash Your Storage Performance by Using NVM Express™ B...Hands-on Lab: How to Unleash Your Storage Performance by Using NVM Express™ B...
Hands-on Lab: How to Unleash Your Storage Performance by Using NVM Express™ B...
Odinot Stanislas7.3K views
Accelerating Real Time Applications on Heterogeneous Platforms by IJMER
Accelerating Real Time Applications on Heterogeneous PlatformsAccelerating Real Time Applications on Heterogeneous Platforms
Accelerating Real Time Applications on Heterogeneous Platforms
IJMER383 views
Memory and Performance Isolation for a Multi-tenant Function-based Data-plane by AJAY KHARAT
Memory and Performance Isolation for a Multi-tenant Function-based Data-planeMemory and Performance Isolation for a Multi-tenant Function-based Data-plane
Memory and Performance Isolation for a Multi-tenant Function-based Data-plane
AJAY KHARAT31 views
Intel new processors by zaid_b
Intel new processorsIntel new processors
Intel new processors
zaid_b2.4K views
참여기관_발표자료-국민대학교 201301 정기회의 by DzH QWuynh
참여기관_발표자료-국민대학교 201301 정기회의참여기관_발표자료-국민대학교 201301 정기회의
참여기관_발표자료-국민대학교 201301 정기회의
DzH QWuynh224 views

Recently uploaded

Evaluation of Quality of Experience of ABR Schemes in Gaming Stream by
Evaluation of Quality of Experience of ABR Schemes in Gaming StreamEvaluation of Quality of Experience of ABR Schemes in Gaming Stream
Evaluation of Quality of Experience of ABR Schemes in Gaming StreamAlpen-Adria-Universität
38 views34 slides
Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ... by
Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ...Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ...
Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ...ShapeBlue
129 views10 slides
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue by
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlueWhat’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlueShapeBlue
265 views23 slides
TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f... by
TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f...TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f...
TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f...TrustArc
176 views29 slides
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha... by
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...ShapeBlue
183 views18 slides
2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlue by
2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlue2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlue
2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlueShapeBlue
152 views23 slides

Recently uploaded(20)

Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ... by ShapeBlue
Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ...Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ...
Live Demo Showcase: Unveiling Dell PowerFlex’s IaaS Capabilities with Apache ...
ShapeBlue129 views
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue by ShapeBlue
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlueWhat’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue
What’s New in CloudStack 4.19 - Abhishek Kumar - ShapeBlue
ShapeBlue265 views
TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f... by TrustArc
TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f...TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f...
TrustArc Webinar - Managing Online Tracking Technology Vendors_ A Checklist f...
TrustArc176 views
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha... by ShapeBlue
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...
Mitigating Common CloudStack Instance Deployment Failures - Jithin Raju - Sha...
ShapeBlue183 views
2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlue by ShapeBlue
2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlue2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlue
2FA and OAuth2 in CloudStack - Andrija Panić - ShapeBlue
ShapeBlue152 views
Business Analyst Series 2023 - Week 4 Session 7 by DianaGray10
Business Analyst Series 2023 -  Week 4 Session 7Business Analyst Series 2023 -  Week 4 Session 7
Business Analyst Series 2023 - Week 4 Session 7
DianaGray10146 views
The Role of Patterns in the Era of Large Language Models by Yunyao Li
The Role of Patterns in the Era of Large Language ModelsThe Role of Patterns in the Era of Large Language Models
The Role of Patterns in the Era of Large Language Models
Yunyao Li91 views
"Running students' code in isolation. The hard way", Yurii Holiuk by Fwdays
"Running students' code in isolation. The hard way", Yurii Holiuk "Running students' code in isolation. The hard way", Yurii Holiuk
"Running students' code in isolation. The hard way", Yurii Holiuk
Fwdays36 views
Initiating and Advancing Your Strategic GIS Governance Strategy by Safe Software
Initiating and Advancing Your Strategic GIS Governance StrategyInitiating and Advancing Your Strategic GIS Governance Strategy
Initiating and Advancing Your Strategic GIS Governance Strategy
Safe Software184 views
Redefining the book supply chain: A glimpse into the future - Tech Forum 2023 by BookNet Canada
Redefining the book supply chain: A glimpse into the future - Tech Forum 2023Redefining the book supply chain: A glimpse into the future - Tech Forum 2023
Redefining the book supply chain: A glimpse into the future - Tech Forum 2023
BookNet Canada44 views
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading... by The Digital Insurer
Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...Webinar : Desperately Seeking Transformation - Part 2:  Insights from leading...
Webinar : Desperately Seeking Transformation - Part 2: Insights from leading...
Hypervisor Agnostic DRS in CloudStack - Brief overview & demo - Vishesh Jinda... by ShapeBlue
Hypervisor Agnostic DRS in CloudStack - Brief overview & demo - Vishesh Jinda...Hypervisor Agnostic DRS in CloudStack - Brief overview & demo - Vishesh Jinda...
Hypervisor Agnostic DRS in CloudStack - Brief overview & demo - Vishesh Jinda...
ShapeBlue164 views
Transcript: Redefining the book supply chain: A glimpse into the future - Tec... by BookNet Canada
Transcript: Redefining the book supply chain: A glimpse into the future - Tec...Transcript: Redefining the book supply chain: A glimpse into the future - Tec...
Transcript: Redefining the book supply chain: A glimpse into the future - Tec...
BookNet Canada41 views
Why and How CloudStack at weSystems - Stephan Bienek - weSystems by ShapeBlue
Why and How CloudStack at weSystems - Stephan Bienek - weSystemsWhy and How CloudStack at weSystems - Stephan Bienek - weSystems
Why and How CloudStack at weSystems - Stephan Bienek - weSystems
ShapeBlue247 views
CloudStack Object Storage - An Introduction - Vladimir Petrov - ShapeBlue by ShapeBlue
CloudStack Object Storage - An Introduction - Vladimir Petrov - ShapeBlueCloudStack Object Storage - An Introduction - Vladimir Petrov - ShapeBlue
CloudStack Object Storage - An Introduction - Vladimir Petrov - ShapeBlue
ShapeBlue139 views
Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And... by ShapeBlue
Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And...Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And...
Enabling DPU Hardware Accelerators in XCP-ng Cloud Platform Environment - And...
ShapeBlue108 views
KVM Security Groups Under the Hood - Wido den Hollander - Your.Online by ShapeBlue
KVM Security Groups Under the Hood - Wido den Hollander - Your.OnlineKVM Security Groups Under the Hood - Wido den Hollander - Your.Online
KVM Security Groups Under the Hood - Wido den Hollander - Your.Online
ShapeBlue225 views
Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P... by ShapeBlue
Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P...Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P...
Developments to CloudStack’s SDN ecosystem: Integration with VMWare NSX 4 - P...
ShapeBlue196 views

A Buffering Approach to Manage I/O in a Normalized Cross-Correlation Earthquake Detection Code for Large Seismic Datasets

  • 1. A Buffering Approach to Manage I/O in a Normalized Cross-Correlation Earthquake Detection Code for Large Seismic Datasets Dawei Mu, Pietro Cicotti, Yifeng Cui, Enjui Lee, Po Chen
  • 2. Outlines 1. Introduction of cuNCC code 2. Realistic Application 3. Performance Analysis 4. Memory Buffer Approach and I/O Analysis 5. Future Work
  • 3. 1. Introduction of cuNCC what is cuNCC ? CUDA based software designed to calculate the normalized cross- correlation coefficient (NCC) between a collection of selected template waveforms and the continuous waveform recordings of seismic instruments to evaluate the waveform similarity among the waveforms and/or the relative travel-time differences. Feb 05, 2016 M6.6
 Meinong aftershocks detection • more uncatalogued aftershocks
 were detected • contributed to earthquake 
 location detection and earthquake 
 source parameters estimation
  • 4. 2. Realistic Application M6.6 Meinong aftershocks 
 hypocenter re-location. •traditional method using short term 
 long term to detect events and using 
 1-D model to locate hypocenter. fewer
 aftershock detected, the result contains
 less information due to inaccuracy. •3-D waveform template matching detect
 events, using 3D model and waveform 
 travel-time differences to re-locate the 
 hypocenter, the result shows more 
 events and more clustered hypocenters, 
 which give us detailed fault geometry •over 4 trillion NCC calculations involved
  • 5. 3. Performance Analysis optimization scheme • The cuNCC is bounded by 
 memory bandwidth • The constant memory is used to 
 stack multiple events into 
 single computational kernel • The shared memory is used to improve the memory bandwidth 
 utilization 1. Compute, Bandwidth, or Latency Bound The first step in analyzing an individual kernel is to determine if the performance of the kernel is bounded by com bandwidth, or instruction/memory latency. The results below indicate that the performance of kernel "cuNCC_04 limited by memory bandwidth. You should first examine the information in the "Memory Bandwidth" section to d limiting performance. 1.1. Kernel Performance Is Bound By Memory Bandwidth For device "GeForce GTX 980" the kernel's compute utilization is significantly lower than its memory utilization levels indicate that the performance of the kernel is most likely being limited by the memory system. For this kern factor in the memory system is the bandwidth of the Shared memory.
  • 6. 3. Performance Analysis performance benchmark This cuNCC code can achieve high performance without high-end hardware or expensive clusters, optimized CPU based NCC code needs 21 hours for one E7-8867 CPU (all 18 cores) to finish mentioned example while a NVIDIA GTX980 only costs 53 minutes. Hardware Runtime (ms) SP FLOP (×1e11) Achieved GFLOPS Max GFLOPS Achieved
 Percentage Speedup E7-8867
 (18 cores) 2968 1.23 41.36 237.6 17.4% 1.0x C2075
 (Fermi) 495 1.8 363.83 1030 35.3% 6.0x GTX980
 (Maxwell) 116 1.8 1552.80 5000 31.0% 25.6x M40
 (Maxwell) 115 1.8 1569.86 7000 22.4% 25.8x P100
 (Pascal) 62 1.8 2911.84 10600 27.5% 47.9x
  • 7. 4. Memory Buffer Approach and I/O Analysis the I/O bottleneck After improving the computational 
 performance with GPU acceleration, 
 I/O efficiency became the new 
 bottleneck of the cuNCC’s overall 
 performance. The output file of cuNCC is an 1-D vector of similarity coefficients saved in binary format, which size is equal to seismic data file. CPU NCC I/O operations cost roughly 10% of the total runtime, while the GPU code I/O cost more than 75% of total runtime. 0 125 250 375 500 NCC(CPU) cuNCC I/O Compute
  • 8. 4. Memory Buffer Approach and I/O Analysis test environment The SGI UV300 system has 8 sockets 18 core Intel Xeon E7-8867 V4 processors 16 DDR4 32GB DRAM run at 1600 MHz. 4 TB of DRAM in Non-Uniform Memory Access (NUMA) via SGI’s NUMALink technology. 4x PCIe Intel flash cards for a total of 8TB configured as a RAID 0 device and mounted as “/scratch” with a 975.22 MB/s achieved I/O bandwidth (with IOR). 2x 400GB Intel SSDs configured as a RAID 1 device and mounted as “/home” with a 370.07 MB/s achieved I/O bandwidth. The software we used were GCC-4.4.7 and CUDA-7.5 along with MPI package MPICH-3.2.
  • 9. 4. Memory Buffer Approach and I/O Analysis use CPU memory as a buffer •Most GPU-enabled computers have more CPU
 memory than GPU memory. 
 (in our case 48GB << 4TB) •Fixed data chuck size (120 days’) 
 with different total workloads •on ”/scratch” partition, for every data size, the 
 buffering technique costs more overall runtime 
 than the no-buffering •on ”/home” partition, buffering starts to help 
 after reaching the 2400-day’s total workload •the high I/O bandwidth filesystem, the 
 improvement brought by the buffering cannot 
 cover up the overhead of the memory transfer.
  • 10. 4. Memory Buffer Approach and I/O Analysis use shared memory virtual 
 filesystem as a buffer •we set 2 TB of DRAM as a shared memory 
 virtual filesystem, and measured the I/O 
 bandwidth achieved 2228.05 MB/s. •on the ”/dev/shm” partition, the high 
 bandwidth of shared memory improves 
 performance greatly by reducing the 
 time used for output. •we gathered the runtime result without 
 buffering scheme from all three storage 
 partitions, and the shared memory 
 partition obtains the best performance.
  • 11. 4. Memory Buffer Approach and I/O Analysis I/O test conclusion •for machines support shared memory virtual filesystem, we recommend to use the shared memory as buffer to output cuNCC result, especially when the similarity coefficients are the median result for the following computation. •for those machines do not have shared memory with high bandwidth I/O device, we recommend to directly output the result to storage without the buffering scheme. •for those machines do not support shared memory with low bandwidth I/O device, we should consider to use CPU memory as a buffer to reduce disk access frequency.
  • 12. 5. Future Work • further optimize the cuNCC code on the Pascal GPU platform. • implement our cuNCC code with “SEISM-IO” library, which interface allows user to switch among “MPI-IO”, “PHDF5”, “NETCDF4”, and “ADIOS” as low level I/O libraries.
  • 13. Thank you for your time !